More than four and a half billion gallons of water are lost in transit every day in the United States, according to the U.S. Geological Survey — and 25 to 35 percent of water is lost every year worldwide, according to the World Bank.
Now a young Israeli company called TaKaDu has developed an innovative solution to this multibillion dollar problem.
Water loss costs the U.S. $2.6 billion every year, and the world $14 billion. When you add in insurance premiums for coverage against the damage that result from leaks and pipe bursts, these costs increase substantially.
“That water has been moved, it has been treated, [and] as it leaks, it might be causing property damage,” said Peter Williams, chief technology officer for the Big Green Innovations group at IBM. “The economic value of [preventing leaks] is compelling.”
As water networks have aged, the challenge has grown. TaKaDu’s answer to this problem is to use the data that water utilities already collect. It claims to have developed a complex and proprietary mathematical algorithm that can be used to predict and detect leaks, bursts, inefficiencies and equipment failures — enabling utilities to quickly prevent and respond to breaks and minimize water loss and repair costs. Infrastructure monitoring therefore enables utilities to make the most of their service and pipe replacement budgets. Here’s what its web-based dashboard looks like:
“Utilities already have meters all over their networks, transmitting data all the time,” said Amir Peleg, TakaDu’s founder and CEO. “They measure water flow, pressure and quality. I asked several utilities what they do with the data and they said almost nothing. They just log the data. They use it for operational needs, not for early detection.”
TaKaDu, however, uses this information, combined with data gleaned from other sources, to identify local and regional consumption baselines, as well as daily and weekly use patterns. Common data sources include utilities’ supervisory systems, which keep tabs on water turbidity, acidity and temperature, as well as public weather services.
“TaKaDu is the first company to take all that data and analyze it and provide information to the utility,” said Heather Landis, an analyst at cleantech-focused Lux Research. “They are operating in their own space.”
Last summer, TaKaDu raised $3.5 million from Israeli venture capital firms Gemini Israel Funds and Giza Venture Capital. Peleg then assembled a team of researchers with advanced degrees in computer science and mathematics — many of whom, like Peleg, are garduates of Talpiot, the Israeli military’s elite science and technology training program. Together, they spent a year formulating and tweaking the algorithms that are now the basis for TaKaDu’s water monitoring platform.
These algorithms borrow from network theory and machine learning to identify correlations between the many variables involved in water distribution. To illustrate, Peleg offered a simple example based on readings from meters in two different neighborhoods served by the same utility.
“Imagine there is always a 20 percent difference between meter readings,” Peleg said. “Why? Because one is in a bigger neighborhood. If the gap grows, chances are one has a leak, or it is a problem of faulty meters. If you correlate among many meters, it is quite a powerful mechanism to learn what normal behavior looks like and where you have an anomaly.”
TaKaDu is operating on a software-as-a-service model. Water utilities can provide their network data to the startup via any web browser. The company’s servers then analyze the data in real time to identify events and abnormalities. But rather than selling the software directly to the utilities, which are notoriously risk-averse, Peleg and his crew are partnering with companies like IBM and Schneider Electric, which already sell products to tens of thousands of water utilities around the world. TaKaDu’s blog just posted an interview with Pascal Bonnefoi, the water segment director for Schneider.
“We like [TaKaDu’s] technology and we like their business model,” said IBM’s Williams. “It has very low barriers to entry. It doesn’t require the insertion of equipment in the ground, which is always expensive. We think of it as an excellent complement to what we sell.”
Already, the startup has several paying utility customers in the U.K., the Netherlands, Australia and Israel. Each of these customers pays a monthly fee of $10,000 to $150,000, depending on the scale of their water networks. Right now, TaKaDu counts Thames Water, the United Kingdom’s largest water utility, and Hagihon, Jerusalem’s water utility, among its clients. Both started paying for the service last quarter.
The company is also conducting ongoing trials across Europe, the Asia-Pacific region and the United States. At the Israel Cleantech Summit hosted by the California Israel Chamber of Commerce, TaKaDu generated considerable buzz. Peleg, who sold his previous company — online behavioral targeting and advertising firm YaData — to Microsoft, has since been hailed for his creative thinking.
“It is exciting to see entrepreneurs who have been successful with technology in a different sector apply the ideas behind it to water,” said Rachel Sheinbein, a water expert at CMEA Capital, a Silicon Valley venture firm. “I like the entrepreneur. I also like that TaKaDu can create potential from [existing] data. We will continue to see [the company] grow.”
Such high praise for a new water play is rare in the venture capital community, which has been slow to invest in water entrepreneurship. Less than 3 percent of venture investments in 2009 went into water, according to the Cleantech Group. But TaKaDu seems to have won over some of the skeptics.
“TaKaDu can very quickly prove ‘Is there a return on investment?’ and shorten the sales cycle,” said Sagi Rubin, a cleantech specialist at private equity firm Virgin Green Fund. “There is scalability in the model and room to expand in terms of coming in with one offering for the municipality, and once in, add[ing] another and another and another. That’s a foundation for a very large business.”
Yoni Cohen is a JD-MBA student at Yale Law School and the Wharton School at the University of Pennsylvania. Born in Israel, he is a former college basketball writer for Fox Sports. Follow Yoni on Twitter at @cohen_yoni.
VB's research team is studying web-personalization... Chime in here, and we’ll share the results.